How To Establish EAT (Expertise, Authority, Reputation)?

[Planet13]

The leaked google review guidelines stressed that human reviewers should examine the EAT factors of a site: Expertise, Authoritativeness, and Trustworthiness.

It gave specific methods to the HUMAN raters on how to do this, such as checking review sites, doing google searches on the author, etc.

But how does the ALGORITHM measure these factors?

Does the algorithm actually take into account reviews from third party sites? Does it try to parse whether the wikipedia entry for that author / entity is "positive" or negative?

While SEO's seem to have joined the EAT Bandwagon, none of them seem to have given a clear definition of what the ALGORITHM would use to determine EAT.

Is negative reputation the new negative SEO?

[Shepherd]

But how does the ALGORITHM measure these factors?

My thought is that it does not, not yet. I would say that's why the have the human raters review site in this manner, so google can take the signals from the sites that score well with the raters and find some way to incorporate them into the algo.

[Planet13]

"...so google can take the signals from the sites that score well with the raters and find some way to incorporate them into the algo."

I think you might be right.

Maybe the algorithm is looking for a co-occurrence of signals; signals that the algo CAN measure that co-occur with signals that are better measured by human raters.

[rish3]

I think that's the general gist. The manual reviewers put sites into buckets...wholly manually.

Then, the algorithm looks at the types of signals that an algorithm is able to see, and find commonalities across sites within the same bucket.

Over simplified, but for example, maybe most of the sites in the "high quality" bucket have a favicon.ico that is unique and not found on other websites. So, the algorithm gives you some bonus points for having a unique favicon.

Or, maybe some percentage of the "very low quality" bucket happen to have a very specific piece of html. Not all the "very low quality" sites have it, but some do. And none of the "medium" or "high quality" websites have this at all. Let's say the html is <meta generator="affliate site generator 2.2" />. So, now, if that's in your html source, you get some negative points.

Neither of those is something that the manual reviewer noticed...they just bucketized the data set. Somewhat similar to how bayesien spam filters work.

I'm sure it's more complex that these examples, but following similar principals for whatever signals (or combinations of signals) where it finds correlations in the buckets.

If you watch the spammy niches, you can see the counter-attack as well. Many payday loan sites now have about-us pages, unique themes with images that aren't anywhere else, phone numbers, addresses, hiring/jobs pages, links from legit local directories, etc.

[Planet13]

@ rish3:

Thanks for the well-thought-out input. Much appreciated. I tend to agree with you.

~~~~

Another thought on why the Algorithm would NOT use various review sites directly in ranking:

The review information is not proprietary to google.

Yes, it may be hard to fake yelp reviews, but at the end of the day, that same info is available to bing, DDG, yandex, and all the other search engines.

If google is going to be spending so much on developing their algorithm, I am guessing they are going to want something proprietary.

[Planet13]

So the next question is...

How do small website owners build up their EAT?

rish3 pointed out some ways that traditional churn-and-burn sites are trying to create EAT. What are ways that more permanent sites can do it better?

[Shepherd]

How do small website owners build up their EAT?

That depends, do you want to "date" google OR "marry" google?

I imagine one could fake enough EAT to date google for a bit, but eventually things would unravel.

To truly have EAT and marry google one is going to have to be a legitimate authority, have legitimate expertise, and actual trust.

Small website owners, well in my mind the first thing I think of is local. Does the owner have actual expertise? A degree? Industry experience? what makes the owner an expert? If it's not the owner, does anyone in the organisation qualify?
Local authority. Gota start somewhere, is the owner the person the local media goes to whenever they report on something related to the owners niche? Here's a good example of an authority here in the U.S.: Michael Baden, ANY time the news is doing a story on a high profile death/autopsy they call this guy, he is the go to authority. And that's reflected in google searches.
Trust, again local, media mentions, local association/chamber/bureau/bbb/trade groups. Networking with related industries.

I consider myself lucky to have been building businesses since before google came along. Admittedly I got a little lazy once google got here because it got so easy but you know what they say: If it's easy everyone will do it.

[EditorialGuy]

How do small website owners build up their EAT?

For starters, they can build "best of breed" (or, at the very least, "among the best of breed") sites in their niches.

In other words, don't think about the signals you're sending to Google; think about the value you're providing for users.

[mrengine]

Establish EAT by creating a product, patenting it and directly retailing it yourself. That worked for us. But for other small businesses in our industry that sell related products, they have all succumbed to Google's fetish with Amazon. For those businesses that did make onsite improvements, it had no impact at all on their search positions - except for some normal up and down flux that can't be attributed to investing in content.

In other words, create something so unique that Google can't monetize it directly themselves or through those companies they have very close relationships with (Amazon).

Alternatively you may be able to establish EAT by getting listed on the NYSE. LOL

[rish3]

Establish EAT by creating a product, patenting it and directly retailing it yourself.

I do see a correlation, in e-commerce, where sites that have product pictures, descriptions, and part numbers that nobody else has...seem to do better. There are ways to do that short of being the sole source of the product.

Correlation != cause though. It could just be those sites also happen to have other quality signals.

[Planet13]

@ EditorialGuy:

"In other words, don't think about the signals you're sending to Google; think about the value you're providing for users."

Yup. I agree.

But at the same time, I need some sort of "assurance" that the hard work I put into making my pages better than any other won't be thwarted by the lack of an EAT signal.

[Planet13]

@ Shepherd:

Does the owner have actual expertise? A degree? Industry experience? what makes the owner an expert? If it's not the owner, does anyone in the organisation qualify?

I understand what you are saying and I think that in most cases that is good (if not for google, at least for users).

But oddly enough, it seems like google has (in the past, at least) completely disregarded this.

Most of my competitors are small, one person or mom and pop sites and I find that most of the text that they use is watered-down wikipedia text - which is bad for us since the wikipedia entries liberally stole OUR text and expertise (they do cite us with a nofollow link, if that matters).

[rish3]

"In other words, don't think about the signals you're sending to Google; think about the value you're providing for users."

Sound advice for building a quality site. However, I am not convinced this has anything at all to do with how well you'll rank in Google. With the context that this is the Google SEO sub-forum, it seems fair game to discuss what google is really looking for.

Two example search queries for today that show just how concerned Google is with actual expertise:

cancer cure - old crap story from fox news, 2 MFA sites, one touting how to "read your poop" for signs of disease

thermite - 3 sites with 9/11 conspiracy theories, wikihow (ugh), about.com (double ugh)

Not sure why anyone thinks that Google has a firm grasp on detecting quality and expertise. They still rely mostly on backlinks.

[Shepherd]

Most of my competitors are small, one person or mom and pop sites...

So, do you think that the human raters would give these competitors a high EAT ranking? I'm guessing no. If they do give low EAT ranking AND google uses that information to improve the algo you may see those competitors lose search ranking in the future.

[Planet13]

"So, do you think that the human raters would give these competitors a high EAT ranking?"

I honestly don't know...

Oddly enough, one competitor made a big jump 9from out of nowhere in to the top).

Their company has been around for about 25 years... and it looks like their website was designed about 25 years ago.

Seriously, the shopping cart they use is no longer supported by the software company that designed it, and it DOESN'T meet PCI compliance (I know because we use the same shopping cart software, but they have not updated to the latest release).

They don't really have many spammy links (a few that I could find), and that's about it.

[Saffron]

Not sure why anyone thinks that Google has a firm grasp on detecting quality and expertise. They still rely mostly on backlinks.

Honestly, I have to agree with this. One site (which has two versions) outranks me almost constantly. Brand sites, the content is okay, but not always as informative as it could be.

One example is an article I have, nice and detailed, with images. It's the kind of topic you really do need images on. They outrank me. Another complicated topic, they outrank me with their 300 word article vs my 1,400 word article. It just seems that Google favour them no matter what. Doesn't matter who has better and more detailed content, this site will outrank me. And as they have two versions (I should note, version 1 is just word for word from a book on the topic, I'm not sure if they're the author, I don't think so), and then site no. 2 comes in second.

[aristotle]

rish3 wrote:
Two example search queries for today that show just how concerned Google is with actual expertise:
cancer cure - old crap story from fox news, 2 MFA sites, one touting how to "read your poop" for signs of disease
thermite - 3 sites with 9/11 conspiracy theories, wikihow (ugh), about.com (double ugh)

rish3 -- Trying to make sense out of Google's search results is a waste of time. Every time you try, you'll just end up even more bewildered than you were before.

[Shepherd]

Well, one can build a lot of Trustworthiness in 25 years.

[EditorialGuy]

So, do you think that the human raters would give these competitors a high EAT ranking?

I think it depends on the topic. If it's a topic like aortic valve stenosis, the National Institutes of Health or the Mayo Clinic is going to have more credibility than Joe Blow does--unless, perhaps, Joe Blow has an M.D. and a string of other credentials after his name. But if it's a topic like dog grooming, bread baking, or hostels in Vientiane, credentials aren't likely to matter as much as the human rater sees on the page. (As the expression goes, "the proof of the pudding is in the eating.")

Not sure why anyone thinks that Google has a firm grasp on detecting quality and expertise.

And I'm not sure why anyone would expect Google's algorithm to be perfect. It is--and always will be--a work in progress.

As I said earlier, "Don't think about the signals you're sending to Google; think about the value you're providing for users." Work on the assumption that Google wants to identify and reward quality, and you're likely to do better over the long haul than you'd do if you spent your time trying to second-guess an imperfect and constantly-evolving algorithm.

[rish3]

And I'm not sure why anyone would expect Google's algorithm to be perfect. It is--and always will be--a work in progress.

I'm not arguing that they should be perfect.

I am arguing to what degree their stated desires (identify/reward quality) even match up at all with what's really happening.

I agree with you, in the general context, building for users is the right thing.

I disagree that it has any direct positive effect on ranking in Google. Money spent on backlinks (or attracting them) is vastly more effective, and if done right, almost a guarantee of organic traffic from G.

Money spent on quality has a much smaller chance of improving rankings. And...I believe it mostly has that effect because the quality attracts backlinks. You can achieve that kind of result more reliably with, for example, well-done link bait (e.g, popularity instead of quality).

Lastly, if you really dig into it, most of the talk, patents, etc, from Google regarding quality signals is NOT about rewarding quality. It's about detecting low-quality, and punishing it. That's not the same as rewarding quality at all. In fact, it's the kind of approach that matches what I see in the SERPS. "Stuff that isn't penalized, ranked roughly in # of backlinks order".

[austtr]

Not sure why anyone thinks that Google has a firm grasp on detecting quality and expertise

Agree 100%. Why?

E expertise > Expert advice or opinion. Skill or knowledge in a particular area.

Every aspect of that involves human deduction and interpretation. It was way beyond the capabilities of a piece of automated software

A authority > An accepted source of expert information or advice

And who exactly makes the determination of what is an "accepted source" and "expert information"? Googles algo would not have a clue and even most human reviewers would argue different points of view.

T trust > Reliance on the integrity, ability and/or character. One in which confidence is placed.

Once again… very aspect of that involves human deduction and interpretation.

Sorry to be blunt but promoting EAT as an achievable objective for webmasters, and that Google's search alog can magically interpret all of the above, is one stop short of being cruel. It's unachievable clap-trap.

By all means try to become a subject expert and produce half decent websites but don't think that alone is going to guarantee you Google traffic. Free traffic from Google is an extremely valuable commodity and Google manages it to their own best commercial advantage… fact of life.

We live in a time of big brand domination in the commercial niches due to their domain authority and backlink profiles... if you can't compete on that scale (who can?) then by all means EAT your heart out, but IMO its unlikely to have any significant impact because, as discussed above, the algo alone cannot interpret any of those indicators.

[FranticFish]

I need some sort of "assurance" that the hard work I put into making my pages better than any other won't be thwarted by the lack of an EAT signal.

My vote would be for links from the sort of sites that don't link out normally. The sort of networking-based link building that Wheel does.

Get your content vouched for by authorities.

[Shepherd]

Get your content vouched for by authorities.

Nice, "peer reviewed", no greater EAT than that.

[Shepherd]

Also, under the "How To Establish EAT" category:

I've just finished setting up a new high end Extended Validation SSL Certificate for our main site. For those that may not have done it before, it's a bit of a process. In order to get the certificate the business, location, phone, ownership all must be verified by multiple third parties. Once complete, when secure pages are loaded in a browser the url bar is turned green and the verified business information is listed.

If I were google, that would be included in the "trust" column of things to look for when evaluating a website.

[EditorialGuy]

If I were google, that [an extended validation SSL certificate] would be included in the "trust" column of things to look for when evaluating a website.

For an e-commerce site, maybe.

For information sites? Not likely. If you're dispensing information, credibility doesn't come from an SSL certificate, it comes from credentials and/or citations (a.k.a. links from reputable sources).

[rish3]

If you're dispensing information, credibility doesn't come from an SSL certificate.

I tend to agree, but Google seems to have decided that it is a ranking signal, for all sites.

John Mueller touches on their reasoning here:

[plus.google.com...]

He gets specific in the comments section:

Some webmasters say they have "just a content site", like a blog, and that doesn't need to be secured....

[Shepherd]

For information sites? Not likely.

Mueller on ssl and "just a content site":

Authentication: How can users trust that the site is really the one it says it is? Imagine you're a content site that gives financial or medical advice. If I operated such a site, I'd really want to tell my readers that the advice they're reading is genuinely mine and not someone else pretending to be me.
On top of these, your users get obvious (and not-so-obvious) benefits.

The credibility does not come from the cert, when talking about an extended validation certificate the credibility comes from the organisation being validated/verified by multiple third parties. It does not make the content more credible, just the likelihood that the content comes from a credible source.

[EditorialGuy]

I tend to agree, but Google seems to have decided that it is a ranking signal, for all sites.

The announcement needs to be kept in perspective. Google said it would be a "lightweight" ranking signal for fewer than 1 percent of search queries.

That could change in the future, but I don't think Sammy SEO (who embraces HTTPS because he thinks it will help him in Google) is going to have an advantage over Trustedauthoritysite dot com just yet.

[Awarn]

Where are these leaked google review guidelines? This sounds VERY much like what I heard from Adwords. They mentioned these buckets. I remember 2 years wanting to use a branded name in the Adwords ad and they disallowed it. I couldn't make any progress and then some woman said I need you to be put in a different bucket and then we can approve the ad. I think that was saying your site has to generate enough traffic they can warrant taking the risk of using a trademarked name. This whole discussion sounds identical to what happened. These buckets are screwed up. I think the buckets are in slots of ten. This is what creates the host crowding. When you a couple companies fighting for the position in the 10 slots one might have better technical signals that carry across the domain so they can achieve 5-7 pages ranking highest in the ten slots. So basically you have a case of the chosen few. If they give a couple slots to Ebay and Amazon you will always see the same companies at the top and that is exactly what I see and have seen for the past 2 years. Page 1 results are basically the same, page 2 (bucket 2) same basic players, page 3, page 4 etc. A totally flawed model. And that also explains why new sites have such an issue being indexed. Then I imagine the use of blogs and mentions in blogs helps make you jump from bucket to bucket. Thus you create an epidemic of scrapebox blogs that really create no true value to the user but it is needed in the game to rank.

[superclown2]

I've just finished setting up a new high end Extended Validation SSL Certificate for our main site

I'd be interested to know how you get on with this. One of Google's main ranking factors is trust so, to me, getting an EVSSL for an important site is a logical step.